Discussion: Version 1.0 - milestones and strategies

[FabianHofmann]

I want to open a discussion about pypsa's roadmap towards its first major version 1.0, gather ideas and wishes on how the package should look like and what it should do. To start, here are some of my suggestions

API Revision

• Add docstrings to all functions, classes and modules
• Establish a harmonized naming convention (eg n for networks, c for component, sn for snapshot)
• Increase test coverage > 90% (Improve test coverage #231, Remove pypsa.stats module and move parts into unit testing #234)
• Move to the black coding style with commit hash in .git-blame-ignore-revs
• Integrate pre-commit hooks for code spell checking and possible other checks
• ...

API Extension

• Integrate linopy as new solver interface
• Possibly drop support for pyomo, remove the modules Remove n.lopf() for pyomo and nomopyomo + HiGHS default solver #884
• Add a clustering module with submodule spatial and temporal with according functions Spatial cluster refactor #673 Draft time-clustering module #665
• ...

API Vision (for dreamers)

• Components are not pure dataframes but python classes which represent containers. For example, n.generators is a GeneratorContainer class or Generators class (an subclass of the ComponentContainer or Components) with various attributes like static for the static dataframe, dynamic for the timedependant dataframes, or even shortcuts, like n.generators.p which would directly return the time-dependent dispatch. These containers could have component specific front-end functions like add_ocgt which has all the technological data for OCGT power plants.
• ...

There is plenty of stuff we could add. Please feel free to extent edit the list. It might also be worthwhile to open a GH project for this.